Data discovery call: DSD x Brainforge
Date: February 10, 2026
Source: Granola
Meeting ID: 04035e9e-e9c9-4e9d-8555-7bdaa64ed514
URL: https://notes.granola.ai/t/04035e9e-e9c9-4e9d-8555-7bdaa64ed514
Participants:
- Uttam Kumaran (Brainforge)
- Shivani (drinklmnt.com)
- Jeff Warren (drinklmnt.com)
- Awaish Kumar (Brainforge)
- Paul (drinklmnt.com)
Transcript
Them: Invoicing kind of anything to our customers, but who are the retailers? So we can run reports on products and package, et cetera, and then when we work with third parties, Third party distributors. Sometimes we can get really detailed information about where they’re selling product to, sometimes we can’t. So I think it really depends on the distributor if they want to share that information of where our product is going. And then some will know what we sold to them. And that’s kind of the last visibility we have on it. So I think that’s probably a high level overview of what we’re looking at. Me: And give us tabs like you’re getting fixed. Reporting out of Encompass already? Them: Yeah. Yeah, they have pretty solid reporting. You can do, like, comparison reports or fusion reports of anything you want. I would say, like, extremely solid reporting. Like, that is the thing that has been such a great sell about it is it is like, so we’re going to add the accounting side. Of encompass most likely element distribution, the self distro portion of the business, so not the selling to someone else who then sells it to stores. But just the part where we sell it to ourselves. Element distribution. And then that is distributed out to stores. All of that accounting and inventory, arap everything. Balance sheet is going to live and encompass financials. And then that will come back to us. So instead of initially, we were thinking maybe we brought that back to stuff like, and then into netsuite, but instead we’re going to try to do it. All wrapped up is the goal and encompass. I mean, maybe if that doesn’t work, even do QuickBooks or some other iteration. But I do like the idea of like, hey, we’re just going to do it as a complete separate financials and then bring in to Snowflake the fixed version of it. Me: Okay? Them: Is one thing to note. So that will be there, and then maybe the other side is like thinking of the data streams that do exist. For us and the self distro side. We will have our inventory the whole way. Because we sell it to ourselves. Me: Yes. Them: We’ll know what’s on hand. We will then have exact records of every store that we sold it into. Just to put a finer point on what Paul was saying about these DSD folks, these other distributors that we’re working with, and all of that just from a financial standpoint. That is sales that are going to the larger business of just drink element. Those sales will go out just like they would to a wholesale customer. Or if we’re shipping direct to Target. The customer would be Big Geyser Distribution. Me: Yeah. Them: And so we ship it to them, we will have their inventory. I don’t know if that data will ever be useful, but we should get for a lot of them, we’ll have some relative information about what they have on hand. Like, we all want to know that. As far as, like, hey, when they’re placing another order. Do we have any concerns? Because we also know that Target, say they’re distributed, Target is really overstocked and some of their directs, and they may be shipping warehouses around. And so we don’t want to let big guys or get overstocked and then have to show this product back, so. That’s a piece of data that may exist. And obviously all of us goes back to our spins data or whatever, our store velocity piece as well. We’re like, yeah. I mean, the end thing is it goes into those stores. Whether or not we know which stores, we do know how fast those stores are moving product. So there is some gray area on either side a little bit more when it’s the self, the dsds, the independent distributors. Is it helpful to talk about the org structure in any way, Shivani, or is it really just, like, the financial flow? No, no, no. I think you’re. I think, like, naming the data sources is the right starting point, actually. And, like, one thing I’ll flag on spins is, like, when I think about the data, that we’re getting from Target and Walmart. We’re getting. Utam, correct me if I’m wrong, we’re getting a point of sales data in addition to, like, what they’re ordering, right? So we have visibility into, like, point of sales data with Target and Walmart through the kind of, like, Emerson data flows that we’re getting and the data that we have. We haven’t been able to access Bin’s data yet. And we actually put out a question to Will, saying, like, what from SPINS would you want us to ingest at one point? Right? Like, what is it that you would want us to ingest from spin’s data? So it seems like you actually have a use case versus like, it was. It wasn’t totally clear. Like, I didn’t feel totally clear what all we want to ingest from spin. So I think even, like, defining that a little bit more will be useful. And then we just need to keep pushing spins and I don’t know who the main contact person is there, but Uttam has now asked multiple times. We’ve, like, signed a contract with them. Like, they just haven’t given us access, like, to ingest the data. So that’s, like, a bit of a blocker. But if you have thoughts on, like, what we should be ingesting from. Spins. That’s a good thread to follow. Okay. Yeah. So workflow, proper demand planning is not based on your spend. It’s not based on your velocities at any of these stores. You will get confused. It is just too hard to work backwards from that angle. The reason why you want it is a double check. So, like you should. The demand planning should be based off of. We think big guys are going to pull much. We think unfi. We think Katie’s going to pull this much. Now, what do I know about these eight UNF IDCs? They’re pretty much only servicing whole foods. Whole Foods has been ripping through product. They send a big order. That makes sense. If we’ve got some way where we can know and double check it. Because I mean for the big broadline ones like when we do have some like before we’re totally self distribute at it everywhere and we’ve we’re working with your k’s or unify. Their demand planning is run by a robot. Just 1 million percent and a person is not paying attention to it. And so you will get crazy orders placed all the time. Me: Yeah. Okay? Yeah, so I saw, like, part of the reason why we look at that data, too, is one thing we did for Medical Cloud is basically trying to get a little bit of sense of critical pos and understanding when that’s going to come in. Them: Cool. Me: So understanding. Yes. Their inventory levels. And if we get a sense of, like, hey, they’re getting what we should expect upos to come in. That’s the sort of stuff that we’d be leveraging someone like dating for the dev helps understand that it’s mainly a big gut check. Them: We’re also sorry. There’s a pin in that. We’re doing all of these independent contract negotiations like that is just starting right now. If we have, like, a North Star, this is how we want data. Like, hey, we want a weekly email and that’s ingestible. If we have a hey, it really needs to be some kind of API link. Me: Yes. Three. Them: Or, like, if you all could be thinking through that piece on your end of, like. Me: Total. So we. We just did this similar. We kind of drafted something for the retailer side. You know, again, in my experience, you’re working with each other sort of like whatever they can give you. So we have more leverage over some of the distributors. And it’s helpful for us to lay out, like, best integration situation. I’m happy to do that. Them: Totally. Me: I totally opinionated on that. Them: That’s awesome, because it really is just like a they are. They have all these things they really care about, and this is when they don’t really care about. But we can bake into the contract from day one and be like, no, this is what we require, and if you are not doing it, you’re in breach of contract, which none of these people are going to want to be a breach of contract with us. Me: Yes. Please. Yes. Okay, great. Them: That’s the whole deal is, like, get these locked in. Me: Ok? Ay. Cool. Them: And hold it for as long as possible. So this is. It’s a really easy leverage situation. Me: Yeah. So if you have, if you have, if, if you at any moment receive from any distributors, like, documentation, the way we can pass this data, that would be helpful. If you haven’t, that’s also fine. We can outline sort of a couple of the methods that we would prefer to receive data and then also somewhat the minimums. And maybe we can drop that together and have that sent over. Them: That probably makes more sense. Like, Paul, I imagine, has some examples in the past. I’m like, Guayaqui data that you’ve gotten from folks. Actually, I don’t know. Paul, would you have some dsd data? I don’t think so. No. We’re just starting these discussions. Like, we literally just went to the people, like, two weeks ago, like the kind of big New York one. So I think we’re in the very much like, the world is always thorough. We ask for what we want. Me: Okay? I can put, like, a simple one. We’ll just put it, like, a one page together of, like, gets you unable to call on somebody, even if it’s acting points. But our setup looks like. We would prefer to get API access. Or, like, at minimum, if they can’t support that, like, what would be requiring. Them: Yeah. Okay? Great. Can you. Sorry. Can you give me a little bit of background? On what? What the goal is here? Yeah, sure. So is that that question for me, Paul, in terms of, like. Yeah, yeah. Okay. So to give you some sense of what we’re trying to do at large with this, like, with this body of work, Right now, a lot of folks across element use data. But go into their individual systems to, like, download, you know, the latest spreadsheet from the latest month and try to make sense of it. And what we’re trying to do at large is build a data stack to say, okay, let’s ingest all of the data into our warehouse at element like, so we’re using Snowflake. I don’t know how familiar you are with elements of a pieces of a data stack, but we’re using Snowflake. So we’ve ingested, let’s say, like, Shopify data. We’ve ingested Shopify’s data and then we’ve ingested Wholesale Team CRM. The Wholesale Team CRM that then Google Sheets. We’ve since been able to join those data sets to then say, like, here, Madison and Laura, like, now you have these clean tables you can reference that have, like, historical data and all these fields that you care about. Every time to, like, try to find that thing every month, right? Like, you now have, like, a live, kind of live data set that’s like, updating daily that you can, like, pull from. I guess, like, zooming out. Is it for demand planning and forecasting? Like, is that the final goal? Like, what? I’d say, like there are many goals to this work, right? It’s like being more data enabled across the business while supply demand is like something further on out. You could imagine there are questions right now. Like, I think when you and I spoke last, correct me if I’m wrong, there was a question like okay, which geographies would we want to go into for self distribution? Right? And right now you can look at maybe like, some version of some data to say, like, okay, how’s retail data working across geographies? But if you just wanted to answer like, how is California or Los Angeles right now across all my channels, like, the goal for this work is that you could then say, I now see a chart by Geo that shows me E Commerce revenue, wholesale revenue, retail revenue and distribution revenue. Like all in in one place. And you can split it by Geo. Right? So, like, yes, like, the end goal, like, you could say, like, is ultimately supporting better supply, demand management. Like, that’s like, one piece of it. But I think just for anybody trying to make a business decision, it’s like, right now people have kind of, like, narrow views of the thing, but we’re trying to, like, stitch it all together, if that makes sense. So Will can say, how is this geography performing across all my channels? And Laura can say, You know, which, which wholesale partners are at risk of turning right now that haven’t, like, ordered in 365 days. And, like, we can come up with different business questions as we clean the tables, as we put things together. And so one layer of this work is like, yes, ingesting the data, cleaning the data. A big part of this work that Brainforge is doing is heavily documenting what each, what each metric means. And once you have like a really nice, like, semantic layer, it’s called, of like, what each, each metric means, we hope that we can have like a BI software that we’re using to then have people like you be able to query things and with, like, natural language, like, kind of like AI you say, I want to see all of Los Angeles split by channel revenue for the last year monthly, or something like that. And then hopefully, if we’ve defined everything really, like, really beautifully, this search that you do will generate you a chart. That, like is helpful for you to, like, answer the question that you need. Got it. Okay. Yeah, that’s helpful. Great. Can you say it again? I’m so fired up to think about. I get kind of hyped, but, like, I. Like I get kind of hyped about this. Me: All we do for that was awesome. I mean. Them: Running through a wall. Me: Part of this is like, you know, another way. I’ve had a really dumb of dab for my team. Is like we’re trying to help people make more decisions and more accurate decisions. And the time spent now in waiting for Data to get there. Or data not being available, and you have to make potentially an accurate decision. It’s like what we’re trying to attack broadly across the business. There are some cases where we have a lot of data. Right. The E Commerce side of the business. The marketing side of the business. People lose tons of digital analytics to do that. There’s also a lot of areas where we’re going to kind of fight to model, like, the retail distribution things where, like, the data may not be as clean. But of course, you guys are making tons of decisions weekly and monthly that we want to support people being a platform. Them: I mean, just to throw a quick example, that’s like, top of mind for me right now. Hey, we’re going to have these monthly reporting things. Like, we need to do all of this, like, actual versus plan. So I’m wondering if there’s some way to, like, have, like, a plan entry. That could go in for, like, hey, this is our roadmap to growth. And it just very broad strokes. Like we have all the routes. How had they been doing over the last month? What’s the difference? What did they, you know, compared to the last three months, what’s the difference in revenue? Like, you know, if we’ve this would not be in there because it’s a team piece. But we might also have like, hey, it broken out by each channel. How are they doing versus the goal for that channel, like their stores. On average right now, we’re going to have like five different reports that we’re going to have to run if we want to use this data. And. Me: And Jeff for solving this today, like, Them: Going into Encompass and having built out reports that give this information. And so you would go to five different pages. Because you can save them. You click them, you’d get it. You put it in here. And then you’d go, okay. I’ve analyzed it. I’m going to highlight the things that are critical. Right now, and then I’m going to type up two paragraphs on it. And that’s what I’m going to do. Every month, over and over. You can build a Fusion report that can run all this for you if you know how to do it. Literally, one report that would do the whole damn thing. Yeah, yeah. Okay, well, that’s why we got to hire John. So question for you, Jeff. Like, I think as we start to think. So first, let me reflect on, like, the. The pacing or the relative to target kind of thing, right? It’s not like Brainforge has ingested our okrs to say, like, okay, like. This is how we are doing relative to how we thought we would be doing. There’s, like, a world in which, when I think about, like, Yes. Just historicals tells you one thing, but, like, how we’re doing relative to where we thought we were going to be is kind of where like, the, the monthly reporting the juice comes out of. Like, what lessons did we learn? And so, like, the. When I think about, like, how to format it, I like to think about it as, like, I have my historicals, and then for my, like, recent data, I’m able to see kind of, like an average, kind of, like, trend relative to where I thought I was going to be. And, like, that’s like, what we’re playing with with Brainforge is like, yes, like, they’re focusing on ingesting. Clean. Ingesting the data, cleaning it up, documenting it. But, like, I’m trying to think about, like, what does that analyst support, whether it be from Brainforge or elsewhere? That helps us kind of, like, format things nicely, show you pacing to plan, and like, that we have a consistent structure for the metrics that matter to us. Like when I. When you just show, like, what you had flashed up a second ago, can you actually pull that up again? Like, curious what kind of metrics we’re looking at here. You had, like, the revenue by route, by root, and then I don’t remember what else you had. Channel Store Average dollars Total number of stores. So each of these is different channel, and then it’s like dollars. And then also, what’s the percentage of target? So if targets 3,000 and this one is at, you know. 2000, you’d be $2,000. 66%. And really only because it would look wonky in a Word document. I think I’m going to move this to a sheet, definitely. We love. It’s helpful just to see it because, like, when I look at this, right? Like, what my brain does is like, okay, what, like your end roots. Like, what. What period of time am I looking at? Like, the hope is that eventually you’re having, like, you have the rows of all the key metrics. That you want. And it’s actually like, we’re able to, like, really populate it with the data that we’re pulling. And then you can see what was it in January, February, March, April. And then you can also show, like, you can layer in how were they relative to target. But when I see charts, like this, or like, sorry, tables. Like this. I’m kind of like. My brain goes, what period of time? What am I thinking about? Like, it’s. It’s kind of hard. You get a feel for it. So I think, like, agreed. Putting it in a nice streamlined format in Excel is preferable. And then we. Can actually hopefully one day ingest the data, feed it to the spreadsheet, and then just kind of have formulas pulling things in. So you’re not even updating things manually at all. Total. Maybe you can do that. On the. Well, I don’t know. Can you put like a budget and I bet you can put a budget encompass. I asked them about that. You can have a budget encompass. I wonder if you could make more of a plan in any event, like. Yeah, seems like there’s, there’s ways to build into it. And yes, that is the idea would be that you would have like, here’s what’s going to happen in a quarter or annually would stack. But, yeah, we’re just trying to give some context to, like, hey, no, this is maybe another use case of where we could get something. Me: Yeah. This is the exact context. Them: Is it super helpful. Yeah. And like, even Jeff, I know that’s like a draft, but if you made a copy of it, we could just like, put it in the folder of, like, the discovery call for DSD for Brainforge, just to reference. These are the kinds of metrics that they care about. Just when they get to that and I if I do the zoom out again in terms of, like, where we are in the process, like, we have started with the commercial side of the business, the revenue side of the business. So we’ve ingested Shopify, we’ve ingested, like, we have like, our retail data for Walmart and Target specifically and and Snowflake. And then we have like, next up, we’re trying to ingest our e commerce data. Like, we’re trying to adjust other e commerce data. We have Shopify, but Amazon, Walmart.com. right. So those are like the things that are next in the queue for us to ingest and then eventually we also want to ingest, like, marketing data so that we’re able to like, like, understand how like, meta ads and things like that layer in to like, the impact on revenue. So all of that is like, in the queue to be ingested. But this is just an initial discovery call with you guys to say, ok, what are the data sources we need to, like, eventually build pipes for? And it just like. Then we’ll add it to the Gantt chart as we focus on all the revenue sources of the business. Well, that’s. I guess that’s helpful. Context and brings me back to the org chart question. There is spend like there are expenses that will come through distribution that our marketing expense. And so there will be like all of our kind of, you know, just on, like, an asset level, there will be, you know, point of sale, display, it’ll be coolers. Like, so there’s kind of that marketing spend. And then there’s also going to be, like, shared advertising expenses, whether that’s out. Of home. Sponsorship type pieces. And then. All of this sampling will be an element expense. So just doubling down, I’m sure whatever you’re doing to monitor the amount of product that we sample. Here and will need to be coded as distro specific sampling and some kind of a manner. But that’s like a huge piece of, like, our whole marketing playbook will be just a ton of products going out. I think the staffing piece will be easy because they’ll just be in the org chart. Me: Yeah. Them: Total as that group. And, like, you know, so it’s funny. It’s like, initially, right now, it’s like revenue focus. It’s like, okay, what’s. What’s coming in. But we had the conversation with Wholesale. Like, we’re, like, kind of, like, trying to tie the bow a little bit. On our work with wholesale just as like a use case. Like, we made some clean tables. We can, like, show you an example. But. But then, you know, they have their similar questions of like, okay, I’ve given this person a fridge. Right. So I’ve given this person a fridge, this wholesale partner, a fridge. Did that increase sales? So that’s like analysis that they’re going to need to be able to do eventually with better data. And I’m hoping the BI tool can help. Like people kind of self serve on that front. But it’s just helpful to know, like, what fields are important to you because there’s so much data out there. So for them, it’s like, who got a fridge versus not when did they get a fridge? And we’re like, ok, if that’s, like, useful in your summary table, we can include that and so you might have similar things like that when you think about your customer table, like, you might want to say, okay, which of these folks got X, Y, Z as a thing for me to track? And thanks. How much money has been pumped into this specific market? How is it impacting velocities and convenience in grocery. Yeah. And, you know, specialty retail. Yep. And that’s like, eventually having, like, your P N L by market kind of thing. Yeah. Okay? That’s helpful. Just. Yeah, go ahead. Well, that we’re going to do. I’m thinking of it more as, like, hey. If you could break it out by, like, maybe that is all going to be P and L based. Yeah, we might be all that with, you know, it’s fine. Okay, great. So Uttam encompass as like is is encompass something that we need to, like, start ingesting into Snowflake. What do you. What are your thoughts after like here? Me: My guidance is like the way you described it, which is like, there will be fixed reports for the compass that we then bring into Snowflake. We have several tools available, like, no reporting happening, so we have to kind of build it all the way up. For example, you consider Shopify and Amazon. Both of those have nuances from the fixed reports, so we actually build them back up from the order levels. So that we’re making sure that they all match. And the way we handle things like refunds and discounts. In this situation. It seems like it’s just one system. That actually just getting access to the final reports and being able to bring those in. It’d be a good way for us to make sure that they end up in Snowflake and the people can repent. And then be vegetable would be our tool. If you could give all the transformation in there. Then assume. Like that’s probably what I’m thinking about right now. Them: I think if you just tell us what you need, we can get it. You know, if it’s inventory or depletions or sales data, like at customer level or channel level or whatever, like, it can all be run. And I think they have all the API integrations. So, I mean, or you can just set up. Manual reports to be sent. Either way, it’s. Yeah, I think it’s all possible. So the. Usually you get your, like, API access and stuff from the. Via the tech team. And so I think this is just something on the list that we can add. And then, and then, like, once you get access to the system, it’s like us figuring out what the best mode of ingestion is probably. Me: Yeah. What exactly we need? Yeah. Them: Because we’re going to come to you. A lot of I think this has to be a two way street on, like, what we actually want you to have from Encompass. Me: Yes. I thought of it. Them: The tech team. Zero involvement in Encompass, by the way. Really, like, in terms of, like, just being able to give API access and stuff like that. There is now. Like, we’re in a huge fight with them right now over the API access situation, so probably. Me: Okay? So I think it’ll only live in Quickbooks. I think it’ll be in Shopify. This is two ways. I mean, either we can say, okay, if QuickBooks has a source of truth revenue, Then just pull the NFL. They can just put don’t revenue for us. And or we get a direct selling compass. I think if we get it from QuickBooks, and we would have, we would need that either way. And then we would join in the Encompass data to that for more granularity. Okay? Yeah. Y. Eah. Yeah. Yeah. So this is going to be Paul Nixon, like what we’re getting from Shopify and then a lot of that holds their application data. This doesn’t Google sheet. We’re combining them and doesn’t update them. Dmv. Okay? Okay? O. Kay. Yeah. Yeah. Like, in terms of. Yeah, I think we just have a total, but we can continue to have different metrics. It’s not that bad. Coming out. If there’s anything that this wholesale sheet can serve. Feel free to use it. And then, yeah, there’s ways that we can split. And to be perfect, I think we should have some information. On it. Have you even been wholesale? Team is looking at them. So that should be the sorts of truth. This. You should. You should be good. We are trying to make it kind of unbreakable, so there’ll be other shoots in it that you can’t touch that start getting sync from Snowflake. Who can drive? I’ll just say, like, we couldn’t do anything. But then we bring it to this table, and then you can filter it and change it. But again, we’ll figure it out. So we’re trying to add some value for the DI tool, and things are there. Thank you. I appreciate it. Thank you. Five.